Transient ischemic attacks (TIA) are critical to identify because prevention therapy can reduce the risk of future vascular events by > 50%. Diagnostic testing and therapeutic intervention must start as soon as possible because 10-25% of TIAs have a stroke within 90 days. Because so many patients present emergently with transient neurological events the large majority of whom do not go on to have a stroke, methods for identifying TIAs at high risk for stroke have been sought so that work up and treatment can be targeted to those who need it most to save time, money and limited resources. Though the ABCD2 score and brain Diffusion Weighted Imaging-MRI (DWI-MRI) have improved prediction of which TIAs have a stroke, their sensitivity and specificity for prediction of individual cases i poor. In this proposal we propose that peripheral blood leukocytes and platelets play pivotal roles in which TIAs go on to have a stroke and by assessing RNA in whole blood we can evaluate leukocyte and platelet function in TIA patients who go on to have stroke versus those that do not have a stroke. We hypothesize that specific coagulation and immune genes are activated in TIA patients that predispose them to have a stroke by 90 days compared to those TIA patients who do NOT have a stroke by 90 days. A subset of these leukocyte and platelet mRNA genes will predict TIAs who have a stroke by 90 days. This hypothesis is addressed by the following specific aims. Aim #1 (Derivation Cohort): Demonstrate that mRNA expression measured using RNAseq from whole blood differs in a derivation cohort of TIAs that go on to have a stroke by 90 days compared to those TIAs who do not have a stroke by 90 days. Demonstrate that most mRNA found to be regulated using RNAseq are also significantly regulated when measured using qRT-PCR. Aim #2 (Derivation Cohort): Apply machine learning algorithms to the mRNA from Aim #1 to derive an optimal subset of mRNA regulated by both RNAseq and qRT-PCR that predict which TIAs have strokes by 90 days compared to those who do not with >95% sensitivity on cross-validation. Aim #3 (Validation Cohort): Use machine/prediction learning algorithms to demonstrate that the genes from Aim #2 when measured using qRT-PCR on an independent validation cohort predict which TIAs have a stroke by 90 days with >85% sensitivity. The goal of these studies is to discover mRNA profiles in blood that predict which TIA patients go on to have strokes by 90 days. When confirmed in future studies, this will direct in depth testing to those high risk TIAs most in need in order to prevent strokes, and decrease unnecessary testing in those with low risk of stroke. Equally as important, the genes discovered to be associated with high risk of stroke in TIA patients will represent potential novel stroke prevention targets.